Ethernet Passive Coaxial Network (EPCN) is a technique based on Ethernet. It adopts a point-to-multipoint structure and adopts a passive coaxial-cable for transmission. In uplink, data flows are transmitted in Ethernet packet bursts.
FIG. 1 is a schematic diagram illustrating a network structure of an EPCN system. As shown in FIG. 1, the EPCN system mainly includes a Coaxial-cable Line Terminal (CLT), a branch/distributor and a plurality of Coaxial-cable Network Units (CNUs). The EPCN system is connected with various external Ethernet devices via the CLT and is connected with user terminals via the CNUs. The EPCN system is applicable for multiple services. And at present, the most popular application is Ethernet broadband to home for users in buildings.
In an EPCN system, each CNU is generally connected with multiple user terminals, e.g., a CNU may be connected simultaneously with an Internet Protocol Television (IPTV) set-top box, an IP phone terminal and a Personal Computer (PC). In practical applications, different user terminals have different QoS requirements. For example, the IPTV set-top box has a very high QoS requirement and the IP phone terminal has a relatively high QoS requirement, while the PC has a relatively low QoS requirement. Thus, in order to meet the QoS requirements of the user terminals, the CNU should be capable of determining QoS levels for packets from different user terminals after receiving the packets so as to process the packets differently. Current CNUs, however, have no such complex QoS identification functions in order to keep low cost. Therefore, the CNUs cannot determine QoS level for a packet according to multi-tuple information of the packet, and thus cannot satisfy different QoS requirements of different user terminals, which lowers the service quality of the EPCN system.